¶

ML3PlatformClient¶

ML3PlatformClient(
   url: str, api_key: str, timeout: int = 60
)

Client class is the single point of interaction with ML cube Platform APIs, it is initialized providing the url and the User api_key. Optionally, you can specify the timeout for the operations. Every operation is performed verifying the API Key and the permissions associated to the User that own that key.

Methods categories¶

There are the following types of methods:

entity creation: create the entity and return its identifier. It is used in the other methods to indicate the entity.
entity update: modify the entity but do not return anything.
entity getters: return a Pydantic BaseModel with the required entity.
entity show: print to the stdout the entity, but they do not return anything.
entity delete: delete the entity
job submission: submit a job on ML cube Platform that will take some time. They return the job identifier that can be used to monitor its state.
job waiters: given a job id wait the until the job is completed

Exceptions¶

The Client class raises only exceptions that are subclasses of SDKClientException. The exception has two fields that you can share with ML cube Support to get help in identifying the problem:

error_code: unique identifier of the error
error_message: message that explain the error

The page is structured in different blocks of methods, one for each entity.

Methods:

.create_company¶

.create_company(
   name: str, address: str, vat: str
)

Create a company for the User, this method works only is the User has not a company yet. After the Company is created the User is the Company Owner.

Args

name : the name of the company
address : the address of the company
vat : the vat of the company

Returns

company_id : str

Raises

CreateCompanyException

.get_company¶

.get_company()

Returns the company of the User

Returns

company : Company

Raises

SDKClientException

.update_company¶

.update_company(
   name: (str|None), address: (str|None), vat: (str|None)
)

Update company information.

Empty values will not be updated.

Allowed Roles

COMPANY_OWNER
COMPANY_ADMIN

Args

name : new the of the company
address : new billing address for the company
vat : new vat of the company

Raises

UpdateCompanyException

.get_all_company_subscription_plans¶

.get_all_company_subscription_plans()

Returns all subscription plans associated with the user company.

Allowed Roles

COMPANY_OWNER
COMPANY_ADMIN

Returns

subscription_plan_list : List[SubscriptionPlanInfo]

Raises

SDKClientException

.get_active_subscription_plan¶

.get_active_subscription_plan()

Returns the current active subscription plan associated with the user company. Returns None if no active subscription plan found.

Allowed Roles

COMPANY_OWNER
COMPANY_ADMIN

Returns

subscription_plan : SubscriptionPlanInfo | None

Raises

SDKClientException

.create_project¶

.create_project(
   name: str, description: (str|None), default_storage_policy: StoragePolicy
)

Create a project inside the company. You don't need to specify the company because a User belongs only to one company and it is retrieved automatically.

Allowed Roles

COMPANY_OWNER
COMPANY_ADMIN

Args

name : the name of the project
description : optional description of the project
default_storage_policy : represents the default policy to use for storing data in ML cube Platform

Returns

project_id : str

Raises

CreateProjectException

.get_projects¶

.get_projects()

Get the list of all projects in the company the User has permissions to view.

Returns

projects_list : List[Project]

Raises

SDKClientException

.get_project¶

.get_project(
   project_id: str
)

Get a project with the given id

Allowed Roles

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : project identifier

Returns

project : Project

Raises

SDKClientException

.update_project¶

.update_project(
   project_id: str, name: (str|None), description: (str|None),
   default_storage_policy: (StoragePolicy|None)
)

Update project details.

Empty values will not be updated.

Allowed Roles

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : project identifier
name : new name of the project
description : new description of the project
default_storage_policy : represents the default policy to use for storing data in ML cube Platform

Returns

project : Project

Raises

UpdateProjectException

.show_projects¶

.show_projects()

Show a list all projects printing to stdout.

Example output:

Project ID                Name
------------------------  ----------
6475f8c9ebac5081e529s63f  my project

.create_task¶

.create_task(
   project_id: str, name: str, tags: list[str], task_type: TaskType,
   data_structure: DataStructure, cost_info: (TaskCostInfoUnion|None) = None,
   optional_target: bool = False, text_language: (TextLanguage|None) = None,
   positive_class: (str|int|bool|None) = None,
   rag_contexts_separator: (str|None) = None
)

Create a task inside the project.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : the identifier of the project
name : the name of the task
tags : a list of tags associated with the task
task_type : the type of the task. See TaskType documentation for more information
data_structure : type of data in the task
cost_info : optional argument that specify the cost information of the task
optional_target : True if the target value in not always available. This changes the behaviour and the detection phase of ML cube Platform that will analyse production data without considering the actual target
text_language : required for NLP tasks, it specifies the language used in the task.
positive_class : required for binary classification tasks, it specifies the positive class of the target.
rag_contexts_separator : Separator used to separate rag contexts from different sources. If missing then only one source exists.

Returns

task_id : str

Raises

CreateTaskException

.update_task¶

.update_task(
   task_id: str, name: (str|None) = None, tags: (list[str]|None) = None,
   cost_info: (TaskCostInfoUnion|None) = None
)

Update task attributes.

None parameters are ignored for the update.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
name : the name of the task
tags : a list of tags associated with the task. To remove all the tags then pass an empty list.
cost_info : optional argument that specify the cost information of the task

Raises

UpdateTaskException

.get_tasks¶

.get_tasks(
   project_id: str
)

Get the list of the Tasks inside the project.

Args

project_id : identifier of the project

Returns

task_list : List[Task]

Raises

SDKClientException

.get_task¶

.get_task(
   task_id: str
)

Get task by id.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : identifier of the task

Returns

task : Task

Raises

SDKClientException

.show_tasks¶

.show_tasks(
   project_id: str
)

Show a list of tasks included in a project to stdout.

Args

project_id : the identifier of a project

Example output:

Task ID                   Name     Type            Status     Status start date
------------------------  -------  --------------  --------   -----------------
6476040d583201813ab4539a  my task  classification  OK         03-02-2023 10:14:06

.create_model¶

.create_model(
   task_id: str, name: str, version: str, with_probabilistic_output: bool,
   metric_name: (ModelMetricName|None),
   preferred_suggestion_type: (SuggestionType|None) = None,
   retraining_cost: float = 0.0, resampled_dataset_size: (int|None) = None
)

Create a model inside the task.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
name : the name of the model
version : the current version of the model
metric_name : performance or error metric associated with the model
retraining_cost : estimated costs in the Task currency to retrain the model. This information is used by the retraining tool to show gain-cost information. Default value is 0.0 meaning that the cost is negligible
preferred_suggestion_type : preferred type of suggestion that will be computed to retrain the model
resampled_dataset_size : size of the resampled dataset that will be proposed to retrain the model note: this parameter is required if preferred_suggestion_type is SuggestionType.RESAMPLED_DATASET

Returns

model_id : str identifier of the created model

Raises

CreateModelException

.get_models¶

.get_models(
   task_id: str
)

Get all models of a task.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : identifier of the task

Returns

models_list : List[Model]

Raises

SDKClientException

.get_model¶

.get_model(
   model_id: str
)

Get model by id.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : identifier of the model

Returns

model : Model

Raises

SDKClientException

.get_model_by_name_and_version¶

.get_model_by_name_and_version(
   task_id: str, model_name: str, model_version: str
)

Get model by name and version.

A Model can have multiple versions according to the updates and retraining done. This method allow to get the Model object by specifying its name and the version tag.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
model_name : the name of the model
model_version : the version of the model

Returns

model : Model

Raises

SDKClientException

.show_models¶

.show_models(
   task_id: str
)

Show a list of models included in a task to stdout.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task

Example output:

Model Id                  Task Id                   Name                    Version    Status           Status start timestamp    Status insert date          Metric Name
------------------------  ------------------------  ----------------------  ---------  ---------------  ------------------------  --------------------------  --------------------
64fecf7d323311ab78f17280  64fecf7c323311ab78f17262  model_local_experiment  v0.0.1     ok                                         2023-09-11 08:27:41.431000  ModelMetricName.RMSE

.get_suggestions_info¶

.get_suggestions_info(
   model_id: str, model_version: str
)

Retrieve suggestions associated with a model.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model
model_version : the version of the model

Returns

suggestion_info_list : List[SuggestionInfo]

Raises

SDKClientException

.show_suggestions¶

.show_suggestions(
   model_id: str, model_version: str
)

Show the list of suggestions associated with a model printing them to stdout.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model
model_version : the version of the model

Example output:

Suggestion Id                     Executed    Timestamp
--------------------------------  ----------  --------------------------
79a8710c351c4b6a9ece7322e153f200  True        2023-08-21 10:54:40.386189

Raises

SDKClientException

.set_model_suggestion_type¶

.set_model_suggestion_type(
   model_id: str, preferred_suggestion_type: SuggestionType,
   resampled_dataset_size: (int|None) = None
)

Set model suggestion type.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the task
preferred_suggestion_type : preferred type of suggestion that will be computed to retrain the model
resampled_dataset_size : size of the resampled dataset that will be proposed to retrain the model note: this parameter is required if preferred_suggestion_type is SuggestionType.RESAMPLED_DATASET

Raises

SetModelSuggestionTypeException

.update_model_version_by_suggestion_id¶

.update_model_version_by_suggestion_id(
   model_id: str, new_model_version: str, suggestion_id: str
)

Update model version by suggestion id. To retrain the Model, ML cube Platform provides importance weights through a SuggestionInfo. After the retraining is completed, you use this method to create the new model version in ML cube Platform. By specifying the suggestion_id, ML cube Platform automatically knows which is the reference data the model is trained on.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model
new_model_version : the new version of the model
suggestion_id : the identifier of the suggestion

Returns

job_id : str job identifier of the pipeline in execution

Raises

UpdateModelVersionException

.update_model_version¶

.update_model_version(
   model_id: str, new_model_version: str
)

Update model version with empty reference data. It is similar to create model but it does not actually create a different Model but a new version of the existent one. After this request, it is possible to upload data with add_historical_data. To start monitoring is required to set the reference of this new model with set_reference request.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model
new_model_version : the new version of the model

Returns

model_id : str job identifier of new model id

Raises

UpdateModelVersionException

.update_model_version_from_time_range¶

.update_model_version_from_time_range(
   model_id: str, new_model_version: str, from_timestamp: float,
   to_timestamp: float
)

Update model version by specifying the time range of uploaded data on ML cube Platform.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model
new_model_version : the new version of the model from_timestamp to_timestamp

Raises

UpdateModelVersionException

Returns the job_id associated to the pipeline

.add_data_schema¶

.add_data_schema(
   task_id: str, data_schema: DataSchema
)

Associate a data schema to a task.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
data_schema : the data schema that characterize your task

Raises

AddDataSchemaException

.get_data_schema¶

.get_data_schema(
   task_id: str
)

Get task data schema

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : identifier of the task

Returns

data_schema : DataSchema

Raises

SDKClientException

.get_data_schema_template¶

.get_data_schema_template(
   task_id: str
)

Get task data schema

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : identifier of the task

Returns

data_schema : DataSchema

Raises

SDKClientException

.show_data_schema¶

.show_data_schema(
   task_id: str
)

Show data schema of associated with a task

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : identifier of the task

Raises

SDKClientException

Example output:

Column name       Role     Type      Nullable
----------------  -------  --------  ----------
sample_id         id       string    False
timestamp         time_id  string    False
sepallength       input    float     False
sepalwidth        input    float     False
petallength       input    float     False
petalwidth        input    float     False
class             target   category  False

.add_historical_data¶

.add_historical_data(
   task_id: str, inputs: Data, target: (Data|None) = None,
   predictions: (list[tuple[str, Data]]|None) = None
)

Add a batch of historical data for the Task.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
inputs : data object that contain input data source. It can be None if you upload other kinds of data
target : data object that contains target data. It can be None if you upload other kinds of data
predictions : list of data objects that contain prediction data. Each element is a tuple with model_id and data object. It can be None if you upload other kinds of data. Predictions are mandatory for RAG tasks while not permitted for the other TaskType.

Returns

job_id : str identifier of the submitted job

Raises

AddHistoricalDataException

.add_target_data¶

.add_target_data(
   task_id: str, target: Data
)

Add target samples for data already uploaded on the Task. This operation is used for Tasks with optional target which is manually labelled. For instance, fter the labelling process (maybe with our Active Learning module) you have a set of labelled samples spread over all the uploaded data. Indeed, they can belong to different data batches (historical or production consistent uploads) and can be a subset of the uploaded data.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
target : data object that contains target data

Returns

job_id : str identifier of the submitted job

Raises

AddHistoricalDataException

.set_model_reference¶

.set_model_reference(
   model_id: str, from_timestamp: float, to_timestamp: float
)

Specify data to use as reference for the model with time range. Data need to be already uploaded on ML cube Platform.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model from_timestamp to_timestamp

Returns

job_id : str identifier of the submitted job

Raises

AddModelReferenceException

.add_production_data¶

.add_production_data(
   task_id: str, inputs: (Data|None) = None, target: (Data|None) = None,
   predictions: (list[tuple[str, Data]]|None) = None
)

Add a batch of production data associated with a given task.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
inputs : data object that contain input data source. It can be None if you upload other kinds of data
target : data object that contains target data. It can be None if you upload other kinds of data
predictions : list of data objects that contain prediction data. Each element is a tuple with model_id and data object. It can be None if you upload other kinds of data

Returns

job_id : str identifier of the submitted job

Raises

AddProductionDataException

.create_kpi¶

.create_kpi(
   project_id: str, name: str
)

Create a KPI.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : the identifier of the project
name : the name of the kpi

Returns

kpi_id : str identifier of the created kpi

Raises

CreateKpiException

.get_kpi¶

.get_kpi(
   kpi_id: str
)

Get kpi by id.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

kpi_id : identifier of the kpi

Returns

kpi : KPI

Raises

SDKClientException

.get_kpis¶

.get_kpis(
   project_id: str
)

Get all kpis of a project.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : identifier of the project

Returns

kpis_list : List[KPI]

Raises

SDKClientException

.show_kpis¶

.show_kpis(
   project_id: str
)

Show the list of KPIs included in a project to stdout.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : the identifier of the project

Example output:

KPI Id                    Project Id                Name                    Status           Status start timestamp    Status insert date
------------------------  ------------------------  ----------------------  ---------------  ------------------------  --------------------------
64fecf7d323311ab78f17280  64fecf7c323311ab78f17262  model_local_experiment  not_initialized                            2023-09-11 08:27:41.431000

.add_kpi_data¶

.add_kpi_data(
   project_id: str, kpi_id: str, kpi: TabularData
)

Add a batch of a given kpi with the given project.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : the identifier of the project
kpi_id : the identifier of the kpi
kpi : data object that contains data source.

Returns

job_id : str identifier of the submitted job

Raises

AddKPIDataException

.compute_retraining_report¶

.compute_retraining_report(
   model_id: str
)

Compute the retraining report for a given model

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model

Returns

job_id : str identifier of the submitted job

Raises

ComputeRetrainingReportException

.get_retraining_report¶

.get_retraining_report(
   model_id: str
)

For a given model id, get the sample weights computed and additional information about them included in the retraining report

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the identifier of the model

Returns

retraining_report : RetrainingReport

Raises

GetRetrainingReportException

.compute_rag_evaluation_report¶

.compute_rag_evaluation_report(
   task_id: str, report_name: str, from_timestamp: float, to_timestamp: float
)

Compute the RAG evaluation report for a given task.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles: - At least PROJECT_EDIT for that project - COMPANY_OWNER - COMPANY_ADMIN

Args

task_id : the identifier of the task
report_name : the name of the report
from_timestamp : start timestamp of the samples to evaluate
to_timestamp : end timestamp of the samples to evaluate

Returns

job_id : str identifier of the submitted job

Raises

ComputeRagEvaluationReportException

.get_rag_evaluation_reports¶

.get_rag_evaluation_reports(
   task_id: str
)

For a given task id, get the computed RAG evaluation reports.

Allowed Roles: - At least PROJECT_VIEW for that project - COMPANY_OWNER - COMPANY_ADMIN

Args

task_id : the identifier of the task

Returns

rag_eval_reports : list[TaskRagEvalReportItem]

.export_rag_evaluation_report¶

.export_rag_evaluation_report(
   report_id: str, folder: str, file_name: str
)

Export a RAG evaluation report to a file.

Allowed Roles: - At least PROJECT_VIEW for that project - COMPANY_OWNER - COMPANY_ADMIN

Args

report_id : the identifier of the report
folder : the folder where the report will be saved
file_name : the name of the file where the report will be saved

Returns

None

.compute_topic_modeling_report¶

.compute_topic_modeling_report(
   task_id: str, report_name: str, from_timestamp: float, to_timestamp: float
)

Compute the topic modeling report for a given task. This functionality is available only for tasks based on text data. The data on which the report is computed is determined by the timestamps.

This request starts an operation pipeline that is executed by ML cube Platform. Thus, the method returns the identifier of the job that you can monitor to know its status and proceed with the other work using the method wait_job_completion(job_id)

Allowed Roles: - At least PROJECT_EDIT for that project - COMPANY_OWNER - COMPANY_ADMIN

Args

task_id : the identifier of the task
report_name : the name of the report
from_timestamp : start timestamp of the samples to evaluate
to_timestamp : end timestamp of the samples to evaluate

Returns

job_id : str identifier of the submitted job

Raises

ComputeTopicModelingReportException

.get_topic_modeling_reports¶

.get_topic_modeling_reports(
   task_id: str
)

For a given task id, get the computed topic modeling reports.

Allowed Roles: - At least PROJECT_VIEW for that project - COMPANY_OWNER - COMPANY_ADMIN

Args

task_id : the identifier of the task

Returns

topic_modeling_reports : list[TaskTopicModelingReportItem]

.get_topic_modeling_report¶

.get_topic_modeling_report(
   report_id: str
)

For a given task id, get the computed topic modeling report for a specific report id.

Allowed Roles: - At least PROJECT_VIEW for that project - COMPANY_OWNER - COMPANY_ADMIN

Args

report_id : the identifier of the report

Returns

topic_modeling_report_detail : TaskTopicModelingReportDetails

.get_monitoring_status¶

.get_monitoring_status(
   task_id: str, monitoring_target: MonitoringTarget,
   monitoring_metric: (MonitoringMetric|None) = None
)

Get the monitoring status of a monitoring target (or metric) in a task.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : the identifier of the task
monitoring_target : the type of monitoring target to get the status
monitoring_metric : the type of monitoring metric to get the status If None, only the monitoring target is considered

Returns

monitoring_quantity_status : MonitoringQuantityStatus

Raises

SDKClientException

.get_jobs¶

.get_jobs(
   project_id: (str|None) = None, task_id: (str|None) = None,
   model_id: (str|None) = None, status: (JobStatus|None) = None,
   job_id: (str|None) = None
)

Get current jobs information. Jobs can be filtered by project_id, task_id, model_id or status.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : the project_id to filter job. If None job of every project will be returned
task_id : the task_id to filter job. If None job of every task will be returned
model_id : the model_id to filter job. If None job of every model will be returned
status : the status to filter job. If None job with every status will be retrieved
job_id : id of the job to filter. If None job with every id will be retrieved

Returns

jobs_list : List[Job]

Raises

SDKClientException

.get_job¶

.get_job(
   job_id: str
)

Get current job information.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

job_id : id of the job to retrieve

Returns

job : Job

Raises

SDKClientException

.show_jobs¶

.show_jobs()

Show current job information to stdout.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Raises

SDKClientException

.get_detection_events¶

.get_detection_events(
   task_id: str
)

Get all detection event of a given task.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : id of the task for which you want to retrieve the detection event

Returns

rules_list : List[DetectionEvent]

Raises

SDKClientException

.get_detection_event_rules¶

.get_detection_event_rules(
   task_id: str
)

Get all detection event rules of a given task.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

task_id : id of the task for which you want to retrieve the detection event rules

Returns

rules_list : List[DetectionEventRule]

Raises

SDKClientException

.get_detection_event_rule¶

.get_detection_event_rule(
   rule_id: str
)

Get a detection event rule by id.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

rule_id : id of the rule

Returns

rule : DetectionEventRule

Raises

SDKClientException

.set_detection_event_user_feedback¶

.set_detection_event_user_feedback(
   detection_id: str, user_feedback: bool
)

Get a detection event rule by id.

Allowed Roles:

At least WORK_ON_PROJECT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

detection_id : id of the detection event
user_feedback : user feedback

Raises

SDKClientException

.create_detection_event_rule¶

.create_detection_event_rule(
   name: str, task_id: str, severity: (DetectionEventSeverity|None),
   detection_event_type: DetectionEventType, monitoring_target: MonitoringTarget,
   actions: list[DetectionEventAction],
   monitoring_metric: (MonitoringMetric|None) = None, model_name: (str|None) = None
)

Create a detection event rule.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

name : the name of the rule
task_id : the id of the task to which the rule belongs. The rule will only respond to detection events generated by this task.
model_name : the name of the model, only required if monitoring_target is set to MODEL.
detection_event_type : the type of detection event that this rule should respond to.
monitoring_target : the type of monitoring target that this rule should respond to.
monitoring_metric : additional metric extracted from monitoring target that is monitored
severity : the level of severity of the detection event that this rule should respond to. None means any severity is matched.
actions : the list of actions to execute, in order, when the conditions of the rule are matched.

Returns

rule_id : str

Raises

CreateDetectionEventRuleException

.update_detection_event_rule¶

.update_detection_event_rule(
   rule_id: str, name: (str|None) = None, model_name: (str|None) = None,
   severity: (DetectionEventSeverity|None) = None,
   detection_event_type: (DetectionEventType|None) = None,
   monitoring_target: (MonitoringTarget|None) = None,
   actions: (list[DetectionEventAction]|None) = None
)

Update a detection event rule.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

rule_id : the id of the rule to update
name : the name of the rule. If None, keeps the existing value.
model_name : the name of the model, only required if monitoring_target is set to MODEL. If None, keeps the existing value.
detection_event_type : the type of detection event that this rule should respond to. If None, keeps the existing value.
monitoring_target : the type of monitoring target that this rule should respond to. If None, keeps the existing value.
severity : the level of severity of the detection event that this rule should respond to. If None, keeps the existing value.
actions : the list of actions to execute, in order, when the conditions of the rule are matched. If None, keeps the existing value.

Raises

CreateDetectionEventRuleException

.delete_detection_event_rule¶

.delete_detection_event_rule(
   rule_id: str
)

Delete a detection event rule by id.

Allowed Roles:

At least PROJECT_EDIT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

rule_id : id of the rule to delete

Raises

SDKClientException

.wait_job_completion¶

.wait_job_completion(
   job_id: str, max_wait_timeout: int = 3000
)

Wait that the ML cube Platform job terminates successfully its execution.

Note that this method stops the execution.

Allowed Roles:

At least PROJECT_VIEW for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

job_id : identifier of the job
max_wait_timeout : maximum amount of seconds to wait before launching JobWaitTimeoutException

Raises

JobWaitTimeoutException when the maximum timeout time is reached
JobNotFoundException when the requested job does not exist
JobFailureException when the requested job is failed

.create_company_user¶

.create_company_user(
   name: str, surname: str, username: str, password: str, email: str,
   company_role: UserCompanyRole
)

Creates a new User in the company.

Allowed Roles:

COMPANY_OWNER

Args

name : name of the user
surname : surname of the user
username : username of the user
password : temporary password for the user. It will change this at the first login
email : email of the user
company_role : role of the user inside the company

Returns

user_id : str

Raises

SDKClientException

.get_company_users¶

.get_company_users()

Returns the list of users in the company.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN
COMPANY_USER

Returns

users_list : List[CompanyUser]

Raises

SDKClientException

.change_user_company_role¶

.change_user_company_role(
   user_id: str, company_role: UserCompanyRole
)

Change the company role of a user in the company.

Allowed Roles:

COMPANY_OWNER: can change the roles of all the Users Users apart from other Admins and the Owner

Args

user_id : the user for which the role is updated
company_role : the new role to assign

Raises

SDKClientException

.show_company_users¶

.show_company_users()

Show company users to stdout.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN
COMPANY_USER

Raises

SDKClientException

.get_user_projects¶

.get_user_projects(
   user_id: str
)

Returns a list of projects that the user can view.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN
COMPANY_USER

Args

user_id : the user for which you want to see the list

Returns

projects_list : List[Project]

Raises

SDKClientException

.show_user_projects¶

.show_user_projects(
   user_id: str
)

Shows the projects that the user can view to stdout.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN
COMPANY_USER

Args

user_id : the user for which you want to see the list

Raises

SDKClientException

.add_user_project_role¶

.add_user_project_role(
   user_id: str, project_id: str, project_role: UserProjectRole
)

Add a project role to the user for the given project.

The User Project role can be assigned only to COMPANY_USER because Admin and Owner already have all the permission over projects.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN

Args

user_id : the user for which you want to see the list
project_id : identifies the project
project_role : the project role to assign

Raises

SDKClientException

.delete_project_role¶

.delete_project_role(
   user_id: str, project_id: str
)

Delete the role of the user for the given project.

The User Project role can be deleted only for COMPANY_USER because Admin and Owner have all the permission over projects.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN

Args

user_id : the user for which you want to see the list
project_id : identifies the project

Raises

SDKClientException

.get_api_keys¶

.get_api_keys()

Returns a list of api keys the user has.

Allowed Roles:

COMPANY_USER

Returns

api_keys_list : List[ApiKey]

Raises

SDKClientException

.show_api_keys¶

.show_api_keys()

Shows the list of api keys the user has to stdout.

Allowed Roles:

COMPANY_USER

Raises

SDKClientException

.create_api_key¶

.create_api_key(
   name: str, expiration_time: ApiKeyExpirationTime
)

Create a new api key for the user

Allowed Roles:

COMPANY_USER

Returns

name : the name of the api key
expiration_time : the expiration time of the api key

Raises

SDKClientException

.delete_api_key¶

.delete_api_key(
   api_key: str
)

Delete the api key of the user

Allowed Roles:

COMPANY_USER

Args

api_key : api key to delete

Raises

SDKClientException

.get_user_api_keys¶

.get_user_api_keys(
   user_id: str
)

Get the list of api keys a user has.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN

Args

user_id : the user to get his api keys

Raises

SDKClientException

.show_user_api_keys¶

.show_user_api_keys(
   user_id: str
)

Shows the list of api keys a user has to stdout.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN

Args

user_id : the user to get his api keys

Raises

SDKClientException

.create_user_api_key¶

.create_user_api_key(
   user_id: str, name: str, expiration_time: ApiKeyExpirationTime
)

Create a new api key for the user.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN

Args

user_id : the user to create a new api key
name : the name of the api key
expiration_time : the expiration time of the api key

Returns

api_key : the new created api key for the user

Raises

SDKClientException

.delete_user_api_key¶

.delete_user_api_key(
   user_id: str, api_key: str
)

Delete the api key of the user

Allowed Roles:

COMPANY_OWNER Admin

Args

user_id : the user to delete an api key
api_key : the api key to delete

Raises

SDKClientException

.change_company_owner¶

.change_company_owner(
   user_id: str
)

Change the company owner role from the requesting user to the other user.

Allowed Roles:

COMPANY_OWNER

Args

user_id : the user that become Company Owner

Raises

SDKClientException

.delete_company_user¶

.delete_company_user(
   user_id: str
)

Delete a user from the company.

Allowed Roles:

COMPANY_OWNER
COMPANY_ADMIN: cannot delete other company admins

Args

user_id : the user to delete

Raises

SDKClientException

.get_integration_credentials¶

.get_integration_credentials(
   credentials_id: str
)

Get the credentials with the given id for 3^rd party service provider integration.

Allowed Roles:

At least WORK_ON_PROJECT for the project where the credentials have been configured
COMPANY_OWNER
COMPANY_ADMIN

Args

credentials_id : id of the integration credentials to retrieve.

Returns

credentials : IntegrationCredentials

Raises

SDKClientException

.get_all_project_integration_credentials¶

.get_all_project_integration_credentials(
   project_id: str
)

Get the list of credentials for 3^rd party service provider integrations that are currently configured in a project.

Allowed Roles:

At least WORK_ON_PROJECT for that project
COMPANY_OWNER
COMPANY_ADMIN

Args

project_id : id of the project for which all configured credentials should be retrieved.

Returns

credentials_list : List[IntegrationCredentials]

Raises

SDKClientException

.delete_integration_credentials¶

.delete_integration_credentials(
   credentials_id: str
)

Delete credentials for the integration with a 3^rd party service provider.

Allowed Roles:

At least UPDATE_PROJECT_INFORMATION for the project where the credentials have been configured
COMPANY_OWNER
COMPANY_ADMIN: cannot delete other company admins

Args

credentials_id : id of the integration credentials to delete.

Raises

SDKClientException

.set_integration_credentials_as_default¶

.set_integration_credentials_as_default(
   credentials_id: str
)

Set the credentials with the given id as default for 3^rd party service provider integration.

Allowed Roles:

At least UPDATE_PROJECT_INFORMATION for the project where the credentials have been configured
COMPANY_OWNER
COMPANY_ADMIN

Args

credentials_id : id of the integration credentials to set as default.

Raises

SDKClientException

.create_aws_integration_credentials¶

.create_aws_integration_credentials(
   name: str, default: bool, project_id: str, role_arn: str
)

Create credentials to integrate with AWS. Returns an object that contains the external_id you will need to configure in your trust policy.

Allowed Roles:

At least UPDATE_PROJECT_INFORMATION for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

name : a simple name to identify this set of credentials
default : whether to use these credentials by default when using an AWS integration
project_id : the project in which these credentials will be configured
role_arn : the ARN of the IAM role that will be assumed by ML cube Platform

Returns

credentials : SecretAWSCredentials

Raises

SDKClientException

.create_aws_compatible_integration_credentials¶

.create_aws_compatible_integration_credentials(
   name: str, default: bool, project_id: str, access_key_id: str,
   secret_access_key: str, endpoint_url: (str|None) = None
)

Create credentials to integrate with AWS-compatible services.

Allowed Roles:

At least UPDATE_PROJECT_INFORMATION for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

name : a simple name to identify this set of credentials
default : whether to use these credentials by default when using an AWS integration
project_id : the project in which these credentials will be configured
access_key_id : the access key id
secret_access_key : the secret access key
endpoint_url : the endpoint url of the service. If None, AWS itself is used

Returns

credentials : AWSCompatibleCredentials

Raises

SDKClientException

.create_gcp_integration_credentials¶

.create_gcp_integration_credentials(
   name: str, default: bool, project_id: str, service_account_info_json: str
)

Create credentials to integrate with GCP.

Allowed Roles:

At least UPDATE_PROJECT_INFORMATION for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

name : a simple name to identify this set of credentials
default : whether to use these credentials by default when using an AWS integration
project_id : the project in which these credentials will be configured
service_account_info_json : the json-encoded string containing the key of the service account

Returns

credentials : GCPCredentials

Raises

SDKClientException

.create_azure_integration_credentials¶

.create_azure_integration_credentials(
   name: str, default: bool, project_id: str,
   service_principal_credentials_json: str
)

Create credentials to integrate with Azure.

Allowed Roles:

At least UPDATE_PROJECT_INFORMATION for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

name : a simple name to identify this set of credentials
default : whether to use these credentials by default when using an AWS integration
project_id : the project in which these credentials will be configured
service_principal_credentials_json : the json-encoded string containing the credentials of the service principal

Returns

credentials : AzureCredentials

Raises

SDKClientException

.set_retrain_trigger¶

.set_retrain_trigger(
   model_id: str, trigger: (RetrainTrigger|None)
)

Set the retrain trigger for a given model.

Allowed Roles:

At least WORK_ON_PROJECT for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the id of the model
trigger : the trigger to set. If you want to remove the trigger, set it to None

Raises

SDKClientException

.test_retrain_trigger¶

.test_retrain_trigger(
   model_id: str, trigger: RetrainTrigger
)

Test the retrain trigger for a given model.

Allowed Roles:

At least WORK_ON_PROJECT for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the id of the model
trigger : the trigger to test

Raises

SDKClientException

.retrain_model¶

.retrain_model(
   model_id: str
)

Retrain a model via the configured retrain trigger.

Allowed Roles:

At least WORK_ON_PROJECT for the project
COMPANY_OWNER
COMPANY_ADMIN

Args

model_id : the id of the model

Raises

SDKClientException